| Summary: | In the present study, a new "anti-Bredt" bicyclic urea, 1,3-diazabicyclo[3·3·1]nonan-2-one, and an isomeric bond-bridged reference compound 1,8-diazabicyclo[4·3·0]nonan-9-one, were synthesized. The [3·3·1] urea was prepared in two ways. 3-Aminomethylpiperidine was reacted with phosgene. Deprotonation of the precipitated ammonium salt with triethylamine gave the urea. In another method, thermal depolymerization (in vacuo) of a glassy product arising from the reaction of 3-aminomethylpiperidine with diphenyl carbonate afforded the urea. The [4·3·0] urea was obtained through the reaction of diphenyl carbonate with 2-aminomethylpiperidine. No glassy intermediate was observed. The [4·3·0] and [3·3·1] ureas resisted hydrolysis. The former would not polymerize either in bulk or in solution at any temperature with or without initiator catalysts. At 120° and above the [3·3·1] urea bulk polymerized within 30 minutes with or without catalysts. The uncatalysed thermal polymerization gave lower molecular weight polymers in lower yield than the catalysed one. Cationic initiator, phenylphosphonic acid, and anionic potassium tertiary butoxide catalysts gave highest molecular weight polymers. Comparatively, the anionic initiator generated a higher molecular weight polymer than the cationic one. Dibutyltin oxide which is a coordination metal catalyst was also effective in generating high polymer but the yield compared favorably only with the uncatalysed thermal polymerization. The bulk polymerization of the [3·3·1] urea is one of the rare cases of polymerization of a bicyclo [3·3·1] nonan derivative. The mechanism proposed for the thermal polymerization of [3·3·1] urea involved the dissociation of the urea to an aminoisocyanate which then propagated the polymerization. The reaction of 3-aminopiperidine with diphenyl carbonate afforded a urea which was proposed to be a dimer of the desired urea, 1,6-diazabicyclo[3·2·1]octan-7-one. The unsuccessful syntheses of two ureas possessing two nitrogen bridgehead atoms, 1,5-diazabicyclo[3·2·1]octan-8-one, and 1,5-diazabicyclo[3·3·1]nonan-9-one, and another which was seven-membered, 1,3-diazabicyclo[3·2·2]nonan-2-one, were explained in terms of their canonical forms violating Wiseman's rule. However, a nuclear magnetic resonance (NMR) spectrum evidence of the formation of 1,5-diazabicyclo[3·2·1]octan-8-one in the NMR tube was presented.
|
|---|
